Disk 4 no longer installed, but can still see files when select disk

[jang430]

Hi, 1 of my disks, Disk 4 turned defective.  I initially changed it to "Not installed", but I can still see the drive in Main/ Array Devices.  When selecting the drive in Windows File Explorer, I can see the files.  I thought I should remove it from Unraid, before the files get moved to other drives.  So I did.  Up to now, I can see drive in Main/ Array Devices as "Not installed."  I can still see files in Windows File Explorer.  Why is it so?  How do I remove the "Not installed" in the Main/ Array Devices?  Change my Disk 5 to Disk 4?  Change my Disk 6 to Disk 5?  If so, it shows error. 

 

BTW, after removing the drive, the Parity didn't start checking also.

[tdallen]

You should post your diagnostics, but best guess is that unRAID is now emulating the missing Disk 4 using parity information.  The normal procedure for replacing a defective drive would now be for you to put in a new disk and assign it to slot 4.  unRAID would then rebuild Disk 4 using parity onto the new drive.

 

You can also do a New Config to redefine the disks and slots to get rid of the old Disk 4 - but if you do you will wipe out parity and be unable to rebuild Disk 4.  Normally you would only do this if Disk 4 was empty.

 

Since things sound a little confused, it would be a good idea to post diagnostics (Tools - Diagnostics) and the steps you plan to take to address Disk 4 - before taking any permanent action like doing a new config or formatting a drive.

[trurl]

Do not do New Config. Don't change anything about your configuration.

 

Put the drive back in, then post diagnostics so we can see if there is really anything actually wrong with the drive. Connection problems are more frequent than actual drive failures.

[jang430]

tdallen, trurl, I saw somewhere that says emulated.  Though now that I'm looking for it, I cannot find it.  Sadly, I don't have any plans to replace the defective drive.  So Drive 4 will be "Not Installed" for now, while I have Drives 5 and 6 intact.  I was hoping by maybe doing a parity check, it will move these temporarily emulated drive/ files to free spaces on Drives 5 and 6?

 

Trurl, before I removed the drive, I removed it from the array, and did a format, or pre-clear, I forgot.  Then joined it once again on the array, and still gives me the same error.  I forgot what the error showed already, but I understood it to be that It's a defective drive.

 

See attached Diagnostic.

tower-diagnostics-20161129-2017.zip

[Frank1940]

tdallen, trurl, I saw somewhere that says emulated.  Though now that I'm looking for it, I cannot find it.  Sadly, I don't have any plans to replace the defective drive.  So Drive 4 will be "Not Installed" for now, while I have Drives 5 and 6 intact.  I was hoping by maybe doing a parity check, it will move these temporarily emulated drive/ files to free spaces on Drives 5 and 6?

 

STOP!!!!  You are in GRAVE danger of losing everything that is on Drive 4.  There is absolutely no way that unRAID will save the data on Drive 4 if you remove it from the array!  IF you insist on removing that drive from the array, you will have to move any data that you want to save manually off of that drive using a disk-to-disk copy.  This has to be done while its contents are being emulated which is the state that your array is in right now.

 

Many people would use the built-in  mc    file manger program that is a part of the basic unRAID OS.  Remember that all copying has to be disk-to-disk.  You can't NOT do disk-to-array copying without causing serious data corruption.  (You can do it over the network but you have to enable disk shares which is disable by default because of the potential data corruption problem previously referred to!)

 

[jang430]

Hello Frank1940, noted on the disk to disk.  After moving files off the drive, and removing the drive itself, how do I prevent it from showing Not Installed?

[itimpi]

Hello Frank1940, noted on the disk to disk.  After moving files off the drive, and removing the drive itself, how do I prevent it from showing Not Installed?

the only way to do this is to use Tools->New Config and then re-assign the drives you want to keep.    You will have to rebuild parity after doing this from the drives that you keep. 

 

Before you do this make sure you know what drives are used where.  In particular make sure you do not accidentally assign a data drive as a parity drive as that would wipe its contents.

[Frank1940]

There is a WIKI describing the way:

 

      https://lime-technology.com/wiki/index.php/Shrink_array

 

Read it very carefully and make sure you understand each caveat.  Basically, you will be using the 'New Config' tool in the 'Tools' Menu.  Get a printout of your current array assignments so you have it for a reference so that you can double check that everything is exactly the way you want as you move through each step. 

[jang430]

Thanks Frank1940

 

[trurl]

DO NOT DO NEW CONFIG!!!

 

At least not yet. And certainly not until you have copied all data from the emulated disk.

 

I think you are rushing off on the wrong path when a better one may be available to you.

 

In the beginning of your syslog, there was some fiddling around with disk assignments, but it looks like you've got it close to normal now:

Nov 26 13:08:01 Tower emhttp: SanDisk_Cruzer_Fit_4C530002650219117361-0:0 (sda) 15633376
Nov 26 13:08:01 Tower emhttp: HITACHI_HDS7250SASUN500G_0649KSJUMH_KRVN67ZAKSJUMH (sdb) 488386552
Nov 26 13:08:01 Tower emhttp: HITACHI_HDS7250SASUN500G_0704K7GBUH_KRVN65ZBG7GBUH (sdc) 488386552
Nov 26 13:08:01 Tower kernel: md: import disk0: (sdf) ST2000DL003-9VT166_5YD2XXLP size: 1953514552 
Nov 26 13:08:01 Tower kernel: md: import disk1: (sdg) ST2000DL003-9VT166_5YD2QEX4 size: 1953514552 
Nov 26 13:08:01 Tower kernel: md: import disk2: (sde) ST2000DL003-9VT166_6YD04VWX size: 1953514552 
Nov 26 13:08:01 Tower kernel: md: import disk3: (sdd) WDC_WD20EFRX-68EUZN0_WD-WMC4M1780430 size: 1953514552 
Nov 26 13:08:01 Tower kernel: md: import_slot: 4 empty
Nov 26 13:08:01 Tower kernel: md: import disk5: (sdh) WDC_WD15EARS-00S8B1_WD-WCAVY4375465 size: 1465138552 
Nov 26 13:08:01 Tower kernel: md: import disk6: (sdi) WDC_WD1003FBYX-01Y7B0_WD-WCAW30608687 size: 976762552 
Nov 26 13:08:01 Tower emhttp: import 30 cache device: sdb
Nov 26 13:08:01 Tower emhttp: import 31 cache device: sdc

SMART for those drives looks OK, but there is no SMART for the missing disk4. Do you still have that disk available?

 

I think it is possible, even likely, that there is nothing wrong with that disk and that you could just rebuild to it. That would be the normal way to handle this situation.

 

New Config, while a possible solution, is going to be a lot more effort (and more activity on an unprotected array) if you intend to save that data since it will all have to be copied from the emulated drive to other disks. When reading the emulated drive, unRAID actually has to read all of the other drives plus parity in order to calculate the missing disk's data.

 

So, please answer my question: Do you still have the missing disk available? If so, please put it back in and post another diagnostic so we can check its SMART.

[jang430]

trurl, Great!  Thanks!  I still have it lying around.  Haven't touched it yet.  I'll put it back and post back in here.

[jang430]

Hello all.  I added the unassiged drive back to array, and selected "Start will start Parity-Sync and/or Data-Rebuild. "  Hope this goes well.  I remember previously adding the drive back, but clicking the button on the upper right, using Dynamix Interface.  By selecting Starting Array on the Dynamix Interface, I don't recall it triggering Parity-Sync or Data-Rebuild.  This time around, I started the array using Unraid's default interface.  Will report back how it goes.

[jang430]

Hello.  My disk 4 still shows and "X".  So far, don't know what to do next.  Attached my diagnostics file here.

tower-diagnostics-20161130-1218.zip

[trurl]

Disk 4 was replaced

Nov 30 11:47:36 Tower kernel: md: import disk4: (sdj) ST31500341AS_9VS41JBZ size: 1465138552 
Nov 30 11:47:36 Tower kernel: md: import_slot: 4 replaced

then later begins to throw a lot of write errors

Nov 30 11:56:14 Tower kernel: md: disk4 write error, sector=20245000
Nov 30 11:56:14 Tower kernel: md: disk4 write error, sector=20245008

and it doesn't appear in the smart folder of your diagnostics so it has probably dropped offline.

 

Check connections, and then before you start the array again to attempt to rebuild the disk, post new diagnostics. We still don't know about the SMART of that disk and whether it is good enough for a rebuild.

[jang430]

Thanks trurl, will report back.

[jang430]

Hello trurl, my unraid is set to start array upon startup.  I stopped it, and got the diagnostic file.  See attached please.

 

BTW,  I also used connections from known working drives, and have it interchanged, just to be sure it's not the physical wire.  As you can see here, it's still the same drive that's defective.

tower-diagnostics-20161203-1021.zip

[trurl]

Yes, disk4 has failed. If you don't have another disk to rebuild to then you can go ahead and copy from the emulated drive to other drives then shrink the array by setting a new config without the drive and rebuilding parity, as previously discussed.

[jang430]

Thank you trurl, appreciate helping me look into this

 

Sent from my LG-D855 using Tapatalk

 

 

[poku]

I think I'm in a bit of a similar situation! The disk appears faulty; file system status "-", and red cross besides the drive saying "Device disabled, contents emulated".

 

The error appeared after starting up the server one day, worked just fine when shutting down. Possibly caused by sudden shutdown?

I tried running xfs_repair with -v -L parameters, but it did not do anything. SMART details appear fine. Is it possible that a brand new drive just suddenly fails like this? What can I do next to fix it? Thank you very much in advance!

 

EDIT: I am dumb. If anyone else happens to stumble upon the same question, see the post below for easy fix.

[trurl]

I think I'm in a bit of a similar situation! The disk appears faulty; file system status "-", and red cross besides the drive saying "Device disabled, contents emulated".

 

The error appeared after starting up the server one day, worked just fine when shutting down. Possibly caused by sudden shutdown?

I tried running xfs_repair with -v -L parameters, but it did not do anything. SMART details appear fine. Is it possible that a brand new drive just suddenly fails like this? What can I do next to fix it? Thank you very much in advance!

 

Anyway, here's the diagnostics attached.

SMART for disk1 looks OK. You can rebuild the drive to itself.

 

Interesting, you tried to fix a redball with a filesystem repair. Not appropriate. But we more often see people trying to fix filesystem by rebuilding from parity. Also not appropriate.

 

Fix disabled disks by rebuilding, fix corrupt filesystem with filesystem repair. Very different problems.

[poku]

SMART for disk1 looks OK. You can rebuild the drive to itself.

 

Interesting, you tried to fix a redball with a filesystem repair. Not appropriate. But we more often see people trying to fix filesystem by rebuilding from parity. Also not appropriate.

 

Fix disabled disks by rebuilding, fix corrupt filesystem with filesystem repair. Very different problems.

 

Oh man. I feel dumb now. I knew it should be something easy, but for some reason I missed that part of the wiki.

Does not show faulty anymore, data rebuild in process. Thank you very much!